Search

Nitrogen fixation in root nodules formed by strain 2143 ofBradyrhizobium japonicum andGlycine max (L.) Merr. cv Williams 82 reaches a maximum at 21 to 28 days postinoculation and then begins to decline. The phytoalexin, glyceollin, accumulates in nodules coincident with the decline in nitrogen fixation. Nodules formed by strain 3122, which are unable to fix nitrogen, accumulate even higher levels of glyceollin and do so beginning 21 days postinoculation even though these nodules contain no recoverable bacteria. The typical phytoalexin response occurs within days of infection. The mechanism by which this response in theBradyrhizobium japonicum-soybean combination is delayed 2 to 3 weeks after infection is presently unknown but phytoalexin accumulation could contribute to the inability of the soybean-Bradyrhizobium japonicum combination to maintain high levels of nitrogen fixation throughout the growing season.

A unique method is described for inhibiting nitrogenase. When Clostridium pasteurianum nitrogenase is assayed in the presence of the Mo-Fe protein of Azotobacter vinelandii, all the characteristic activities of nitrogenase are inhibited. C. pasteurianum nitrogenase is unaffected by the Fe protein of A. vinelandii. The Fe protein, but not the Mo-Fe protein of C. pasteurianum, inhibits A. vinelandii nitrogenase. Both inhibitions described result from the formation of an inactive complex of A. vinelandii Mo-Fe protein and C. pasteurianum Fe protein. Complex formation requires active components, as oxygen-denatured proteins are ineffective. The results for titration of components of the complex against each other and kinetic data each indicate that the inactive complex consists of two molecules of C. pasteurianum Fe protein per molecule of A. vinelandii Mo-Fe protein. The results of kinetic experiments suggest that the Fe protein from each organism competes for the same site(s) on the A. vinelandii Mo-Fe protein. The Fe protein of C. pasteurianum will form an active or an inactive complex with the Mo-Fe proteins from six different organisms. Inhibition by nitrogenase components that form inactive complexes provides numeroius ways to study the mechanism of nitrogenase action.

Acetyl-CoA synthetase was purified 800-fold from Bradyrhizobium japonicum bacteroids. A specific activity of 16 mumol/min per mg of protein was achieved, with a 30-40% yield. The purification scheme consisted of only three consecutive chromatography steps. The enzyme has a native Mr of 150,000, estimated by gel-permeation chromatography, and a subunit Mr of 72,000, determined by SDS/polyacrylamide-gel electrophoresis. The optimum pH and temperature are 8.5 and 50 degrees C respectively. The Km values for acetate, CoA and ATP were 146, 202 and 275 microM respectively. The reaction was specific for acetate, as propionate and oleate were used very poorly. Likewise, the enzyme used only ATP, ADP or dATP; AMP, GTP, XTP and UTP could not replace ATP. Acetyl-CoA synthetase showed a broad specificity for metals; MnCl2 could replace MgCl2. In addition, CaCl2 and CoCl2 were approx. 50% as effective as MgCl2, but FeCl3, NiCl2 or ZnCl2 could not effectively substitute for MgCl2. The enzyme may be regulated by NADP+ and pyruvate; no effect was seen of amino acids, glucose catabolites, reduced nicotinamide nucleotides or acetyl-CoA. Inhibition was seen with AMP, PPi, FMN and pyridoxal phosphate, with Ki values of 720, 222, 397 and 1050 microM respectively.

Protein phosphorylation was demonstrated in Bradyrhizobium japonicum bacteroids in vivo and in cultures in vivo and in vitro. Comparison of in vivo-labeled phosphoproteins of bacteroids and of cultured cells showed differences in both the pattern and intensity of labeling. In cultured cells, comparison of the labeling patterns and intensities of in vivo- and in vitro-labeled phosphoproteins showed a number of similarities; however, several phosphoproteins were found only after one of the two labeling conditions. The labeling intensity was time dependent in both in vivo and in vitro assays and was dependent on the presence of magnesium in in vitro assays. Differences in the rates of phosphorylation and dephosphorylation were noted for a number of proteins. The level of incorporation of 32P into protein was only 2% or less of the total phosphate accumulated during the in vivo labeling period. Several isolation and sample preparation procedures resulted in differences in labeling patterns. Phosphatase inhibitors and several potential metabolic effectors had negligible effects on the phosphorylation pattern. There were no significant changes in the phosphorylation patterns of cells cultured on mannitol, acetate, and succinate, although the intensity of the labeling did vary with the carbon source.

A mutant strain of Bradyrhizobium japonicum USDA 110 devoid of [alpha]-ketoglutarate dehydrogenase activity (LSG184) was used to test whether this tricarboxylic acid cycle enzyme is necessary to support nitrogen fixation during symbiosis with soybean (Glycine max). LSG184 formed nodules about 5 d later than the wild-type strain, and the nodules, although otherwise normal in structure, contained many fewer infected host cells than is typical. At 19 d after inoculation cells infected with the mutant strain were only partially filled with bacteroids and showed large accumulations of starch, but by 32 d after inoculation the host cells infected with the mutant appeared normal. The onset of nitrogen fixation was delayed about 15 d for plants inoculated with LSG184, and the rate, on a per nodule fresh weight basis, reached only about 20% of normal. However, because nodules formed by LSG184 contained only about 20% of the normal number of bacteroids, it could be inferred that the mutant, on an individual bacteroid basis, was fixing nitrogen at near wild-type rates. Therefore, the loss of [alpha]-ketoglutarate dehydrogenase in B. japonicum does not prevent the formation or the functioning of nitrogen-fixing bacteroids in soybean.

Adenylate cyclase and cyclic AMP (cAMP) phosphodiesterase have been identified and partially characterized in bacteroids of Bradyrhizobium japonicum 3I1b-143. Adenylate cyclase activity was found in the bacteroid membrane fraction, whereas cAMP phosphodiesterase activity was located in both the membrane and the cytosol. In contrast to other microorganisms, B. japonicum adenylate cyclase remained firmly bound to the membrane during treatment with detergents. Adenylate cyclase was activated four- to fivefold by 0.01% sodium dodecyl sulfate (SDS), whereas other detergents gave only slight activation. SDS had no effect on the membrane-bound cAMP phosphodiesterase but strongly inhibited the soluble enzyme, indicating that the two enzymes are different. All three enzymes were characterized by their kinetic constants, pH optima, and divalent metal ion requirements. With increasing nodule age, adenylate cyclase activity increased, the membrane-bound cAMP phosphodiesterase decreased, and the soluble cAMP phosphodiesterase remained largely unchanged. These results suggest that cAMP plays a role in symbiosis.

The nitrogenase proteins from eight organisms have been highly purified, and a survey of their cross-reactions shows that the nitrogenase proteins from a wide variety of organisms can interact with one another. An active cross-reaction is the complementary functioning of the MoFe protein and the Fe protein from different organisms. Of 64 possible combinations of component proteins, 8 yielded homologous nitrogenases (components from the same organism); 45 of the 56 possible heterologous crosses generated active hybrid nitrogenases; 4 heterologous crosses yielded no measurable nitrogenase activity but did form inactive tight-binding complexes; 6 crosses did not give measurable activity; and 1 cross was not made. All these crosses were assayed for acetylene reduction, and some also were assayed for ammonia formation, hydrogen evolution, and ATP hydrolysis activity. The activity generated by combining two complementary heterologous nitrogenase components depended on pH, component ratio, and protein concentration, the same factors that determine the activity of homologous nitrogenases. However, several crosses showed an unusual dependency on component ratio and protein concentration, and some cross-reactions showed interesting ATP hydrolysis activity.